J . Med. Microbiol. - Vol. 35 (1991), 330-337
01991 The Pathological Society of Great Britain and Ireland
Contact- haemolysin production by entero-invasive Escherichia cok and shigellae KHALEDA HAIDER, M. J. ALBERT, A. HOSSAIN* and S. NAHAR Laboratory Sciences Division, International Centre for Diarrhoea1 Disease Research, GPO Box 128, Dhaka 1000, and Department of Microbiology, University of Dhaka, Dhaka 1000, Bangladesh
Summary. Entero-invasive Escherichia coli (EIEC) and shigellae were tested for contacthaemolysin (CH) with red blood cells (RBCs) of guinea-pig, rabbit, rat, mouse, monkey, man, sheep and chicken; all bacteria showed the best lysis with guinea-pig RBCs. The best culture medium for CH activity of shigellae was tryptic soy broth, and for EIEC it was casamino acid-yeast extract broth with 1 mM CaC12.CH production by all species was best at the slightly alkaline pH which is optimal for growth; it was also dependent on the presence of a large (140-Mda) plasmid. Pre-treatment of bacteria with homologous antisera inhibited CH activity. Various treatments of bacterial cells and RBCs suggested that CH may be a protein molecule, and that a chitotriose-like moiety may serve as CH receptor. RBCs that were incubated with bacteria at 4"C, or with heat-killed bacteria at 37"C, were not lysed; also, isolated cell-surface components (lipopolysaccharideand outer-membrane protein) did not lyse RBCs. This suggests that metabolically active cells are required for CH activity. Production of CH by both EIEC and shigellae is consistent with a common mechanism for the virulence of these organisms.
Introduction
The large virulence-plasmidsof EIEC and all species of Shigella have extensive homologous sequences and are functionally exchangeable.' As the pathogenesis of EIEC and Shigella spp. is similar, this study was initiated to investigate whether CH was also produced by S . boydii and EIEC, other species of Shigella being included for comparison. Factors influencing the production and activity of CH were investigated by the use of virulent and avirulent pairs. Preliminary experiments were also performed to study the chemical nature of CH.
Disease caused by shigellae involves a complex process of invasion and multiplication in the mucosa of the human colon. The process of invasion requires entry of the bacterium into the enterocytes, and lysis of the membranous vacuole which surrounds it after invasion, to allow free multiplication of the pathogen within the cytoplasm. This leads to spread of the organisms to the lamina propria, triggering a severe inflammatory response in the gut mucosa with abscesses and ~ l c e r a t i o n . ~All ~ these features are regulated by a large non-self-conjugative plasmid of 120-140 Mda called the virulence plasmid. Enteroinvasive Escherichia coli (EIEC) produces dysentery like that caused by shigellae; it also possesses a 140Mda plasmid which is associated with vir~lence.~. Sansonetti et a1.6 described a virulence-plasmidmediated, thermoregulated contact-haemolysin (CH) in Shigella, which was active against sheep red blood cells (SRBC). Close contact was required between virulent bacteria and SRBC which was achieved by centrifugation. The production of CH correlated with its capacity to destroy the membranous vacuole surrounding the bacteria after invasion of HeLa cells, and with efficient intracellular multiplication. CH was also demonstrated in S. dysenteriae and S . sonnei strains in another s t ~ d y . ~
Materials and methods Bacterial strains and culture conditions The bacterial strains are shown in table I. Dr D. N. Taylor, Armed Forces Research Institute of Medical Sciences (AFRIMS), Bangkok, kindly provided 26 strains of EIEC from patients with invasive diarrhoea in Thailand. Five similar isolates were from central Australia (2), Bangladesh (2) and Hungary (l).9 The avirulent EIEC strain 2/33 is an isogenic mutant strain of virulent EIEC strain 2, isolated in Hungary ;9 these were kindly supplied by Dr P. Echeverria (AFRIMS, Bangkok). E. coli strain 36 000 was obtained from the Centers for Disease Control, Atlanta, GA, USA. All Shigella strains were isolated from patients at the Clinical Research Centre, International Centre for Diarrhoeal Disease Research, Bangladesh. The CH
Received 19 Dec. 1990; revised version accepted 3 April 1991. 330
CONTACT HAEMOLYSIN OF EIEC AND SHIGELLAE
331
Table I. CH activity and HeLa-cell invasiveness of shigellae and EIEC Number of strains
CH activity (OD,,,)
Number* of bacteria/HeLa cell
Virulent Avirulent Rough
5 5 4
2.19 0.89 0.12
53 0 0
S . sonnei
Virulent Avirulent Rough
5 5 4
1.65 0.12 0.15
54 0 0
S . boydii
Virulent Avirulent
5 5
1.16 0.34
11 0
S.Jlexneri l a lb 2a 2b 3a 3b 4a 5 Y 6 2a
Virulent Virulent Virulent Virulent Virulent Virulent Virulent Virulent Virulent Virulent Avirulent
2 2 5 2 2 2 1 2 2 2 5
1a06 0.72 2.43 0.7 1 1.10 1.14 1.10 2.00 2.27 1.44 0.5 1
25 5 70 3 15 15 12 51 65 20 0
EIEC
Virulent Virulent Virulent Virulent Virulent Virulent Virulent Avirulent
24 2 1 1 1 1 1 1
2.72 1.35 0.87 0.82 0.90 0.56 1.01 0.50
Avirulent
1
0.10
152 15 3 6 6 4 27 0 0
Organism S . dysenteriae type 1
028 029 0143 0144 0164 0124 0136 0143
E.coli 36000
* Each of these numbers is the approximate mean of results from at least three experiments. Virulent strains gave a positive Sereny result, invaded HeLa-cells and bound Congo red dye; avirulent strains did none of these. Virulent strains had a 120-140-Mda virulence plasmid; avirulent strains had either no virulence plasmid or a deleted one, Rough strains autoagglutinated in physiological saline and did not have the virulence plasmid.
activity of bacterial components was studied with S . JEexneri strain 2-61 1, S. dysenteriae type 1 strain Z24623, and EIEC strain 2. All the strains were stored at - 70°C in Trypticase Soy Broth containing yeast extract 0.6% w/v (TSB; Gibco) with glycerine 15%v/v. The growth media were TSB, Casamino Acid-Yeast Extract Broth (Difco) supplemented with 1 mM calcium chloride (CYE) and minimal medium (MM) as described by A hmed et al.
Plasmid DNA was isolated by the method of Birnboim and Doly' and electrophoresed in agarose 0.7% gel.
Assay of CH
Testsfor bacterial virulence
CH activity was measured by the method of Sansonetti et a1.6 Briefly, bacteria were washed once and suspended in phosphate-buffered saline pH 7.3 (PBS) to a concentration of 2 x lo1' cfu/ml; blood was collected in Alsever's solution, washed twice, and the packed RBCs were resuspended in PBS to a concentration of 4 x lo9 RBCs/ml; 5O-pl volumes of bacterial suspension were mixed with equal volumes of RBCs in 96-well microtitration plates, and centrifuged at 2200 g for 10 min to achieve close contact between bacteria and RBCs. The plates were incubated at
Each strain was tested for virulence, both by invasion of HeLa cell monolayers6and by the Sereny test.I3 The latter was performed by inoculating 20 pl of bacterial suspension containing 3 x lo8 cfu/ml into the conjunctival sac of an adult guinea-pig; strains producing keratoconjunctivitis within 72 h were considered virulent. For the former, non-confluent monolayers of HeLa cells on 12-mm cover glass (Kimble, Toledo, OH, USA) in a vial were inoculated with bacteria, centrifuged for 10 min at 2200 g and incubated for 2 h
either 37°C or 4°C for 2 h. Pellets were resuspended in cold PBS, centrifuged at 4°C for 15 min and the OD,,o of the supernates was measured with an ELISA reader (Titertek-Multiscan, Flow Laboratories).
Plasmid analysis
332
K. HAIDER ET AL.
at 37°C. Non-invasive bacteria were washed off with PBS. Monolayers were fixed in methanol and stained with Giemsa dye, and the percentage of infected cells was determined by light-microscopy.To calculate the average number of viable bacteria in each infected HeLa cell, strains were inoculated on to monolayers of HeLa cells in a vial, centrifuged for 10 min at 2200 g, incubated for 15 min, washed three times with PBS, covered with 2ml of Minimum Essential Medium (Gibco) containing gentamicin 200 pg/ml, and reincubated for 2 h at 37"C, Monolayers were then washed, cells were trypsinised and counted, and finally cells were lysed with a solution of sodium deoxycholate 0.5% in distilled water. Dilutions of this suspension were plated to count the number of bacteria/vial. The average number of bacteria in each infected HeLa cell in a vial was calculated as follows : number of viable bacteria number of HeLa cells x percentage of infected HeLa cells
For comparison of the invasiveness of different organisms, the chosen multiplicity of infection was 100 bacteria/HeLa cell. Each strain was tested at least three times; results are presented as the mean from three experiments.
Treatment of erythrocytes Erythrocytes (RBCs) were treated at 37°C for 1 h with various reagents (Sigma): N-acetylneuraminic acid 1% w/v (NANA), lysozyme250 pg/ml, neuraminidase 100-1000 U/ml, and lectins-wheat germ agglutinin 10 pg/ml, phytohaemagglutinin 1.0 pg/ml and Concanavalin A 0.1 pg/ml. Higher concentrations of phytohaemagglutinin or Concanavalin A caused lysis of guinea-pig RBCs.
Treatment of bacteria Bacteria (10gcfu)were treated with: (1) enzymes for 4 h at 37°C-trypsin (Flow Laboratories, Herts) 100 pg/ml, proteinase K (Sigma) 100 pg/ml; (2) rabbit sera for 1 h at 37°C-normal serum and homologous and heterologous antisera against whole bacterial cells;14 (3) heat, at 56°C or 100°C for 1 h.
Extraction of bacterial components Lipopolysaccharide (LPS) of S . JEexneri strain Z61 1, S . dysenteriae type 1 strain 2-24623, and EIEC strain 2 were extracted with hot phenol-water" and purified by ultracentrifugation at 100 000 g for 4 h. Nucleic acid and protein were removed by treatment with DNAase and RNAase, and proteinase K 100 pg/ ml, respectively. Bacterial cells were depleted of LPS by treating cells with edetic acid. Outer-membrane protein (OMP) was prepared by water extraction. LPS and OMP were freeze-dried and reconstituted in PBS. Protein was estimated by d~e-binding'~ and
'
carbohydrate in LPS was estimated by the method of Dubois et aZ.'* CH activity of LPS and OMP was tested at concentrations up to 10 mg/ml. Bacterial cells depleted of OMP and LPS were also assayed for CH activity.
Results InJtuence of growth medium on CH activity Bacteria were grown with shaking in TSB, CYE and MM, and their CH activity was assayed with sheep RBCs (fig. 1). For S.JEexneri,TSB was the best medium, whereas S . dysenteriae and EIEC exhibited maximal activity when grown in CYE; on the other hand, CYE provided only poor CH activity by S . Jlexneri, S. sonnei and S. boydii. CH activity was reduced or undetectable when EIEC and shigellae were grown in static broth or on agar plates of these media.
CH activity on RBCs of different animal species RBCs from guinea-pig, rabbit, rat, mouse, monkey, man (blood group 0), sheep and chicken were compared (fig. 2). All the bacterial species exhibited maximal CH activity against guinea-pig RBCs ;hence, the latter were used for subsequent experiments. Except with S . JEexneri and EIEC, sheep RBCs were poorly lysed. The influence of growth medium was re-tested with guinea-pig RBCs : TSB was the most suitable medium for production of CH by shigellae and CYE for EIEC (fig. 1).
InJtuence ofpH of growth medium on CH activity For S.JEexneri, S . dysenteriae type 1 and S . sonnei, neutral or slightly alkaline pH (7.3-8.5) was suitable for CH production. However, with S . boydii, the activity was reduced as the pH was increased from 7-3 to 9.0, and there was not much reduction in CH activity as the pH decreased from 7-3 to 6.0. EIEC showed a marked optimum at pH 7.3, and this pH was also near the optimal for shigellae (fig. 3).
Inzuence of duration and temperature of growth on CH activity CH activity was examined in bacterial cultures grown for 4-48 h. It was maximal at 6-10 h, but beyond 24 h bacteria exhibited little haemolytic activity. CH activity of both shigellae and EIEC was greatly dependent on growth temperature : strains which were strongly positive when grown at 37°C showed reduced activity when grown at 30°C. Activity could be partially restored by changing the growth temperature to 37°C for 2 h before incubation with RBCs, but
CONTACT HAEMOLYSIN OF EIEC AND SHIGELLAE
333
Fig. 1. Influence of growth medium on CH activity of ShigeZZa spp. and EIEC. Bacteria grown in TSB (m), CYE (R) and MM (m) were tested with sheep RBC. Blank bars (0) to the right of the marked bars indicate CH activity of bacteria grown in the same medium but tested with guinea-pig RBC. OD,,o was measured with an ELISA reader. S. flex = S.jexneri; S. dys 1 = S . dysenteriae type 1.
Fig. 2. CH activity of ShigelZa spp. and EIEC against RBCs of different animal species: guinea-pig, monkey, ;man, W ;sheep, ; chicken, @ and see footnote to fig. 1.
extension of this period to 4 h did not result in further increase of CH activity. No CH activity was shown when the assay was performed at 4"C, even with bacteria grown optimally at 37°C.
.;
rabbit,
m; rat, 1; mouse, @ ;
Association of large virulenceplasmid with C H activity Plasmid profiles of several virulent strains, both shigellaeand EIEC, and of their avirulent counterparts
334
K. HAIDER ET AL.
3-0
2-5
2.0
1.5
1.0
0-5
0 S.flex
S.dys 1
S.sonnei
S.boydii
EIEC
Fig. 3. Influence of pH of growth medium on CH activity of Shigella spp. and EIEC : pH 6.0, TJ ;6.5, IS!;7.3, see footnote to fig. 1.
were analysed to study association of the 120-140Mda plasmid with CH activity. Avirulence was determined by the inability to bind the dye Congored, to produce keratoconjunctivitis in the guineapig,I3 and to invade HeLa cells.6 In all the species, CH activity appeared to be plasmid mediated, because loss of the large plasmid, or reduction in size by deletion, resulted in almost complete absence of CH activity (table I).
''
Correlationof CH activity with invasion potential of bacteria S . flexneri serotypes 2a, 5 and Y, S . dysenteriae type 1, and S . sonnei showed similar invasion potential after incubation for 2 h, reaching 51-70 bacteria/ HeLa-cell (table I), However, EIEC type 0 2 8 showed a much higher degree of invasion (152 bacteria/cell) than shigellae. Other serotypes of EIEC and S.flexneri and strains of S . boydii showed lower invasion capability (3-27 bacteria/cell). Generally, CH activity correlated well with the invasion potential of the organism. However, avirulent strains of S. dysenteriae type 1, although non-invasive in HeLa-cell assay, exhibited slightly higher CH activity than some virulent strains of other ShigeZZa spp. and EIEC. Also, one virulent EIEC strain ( 0 124) exhibited low CH activity although it was moderately invasive (table I).
;8.0,
; 8.5, El;9.0, Ei and
Influence of various treatments of RBCs and bacteria on CH activity Pre-treatment of RBCs with lysozyme resulted in significant reduction (56-75%) in CH activity, but neuraminidase and N ANA had minimal inhibitory effect (table 11). Of the three lectins, wheat germ agglutinin showed 50-67% inhibition of CH activity, whereas phytohaemagglutinin and Concanavalin A had no effect. No reduction in CH activity was noted when bacteria were treated with trypsin, but CH activity was completely lost after treatment with proteinase K, indicating that the CH molecule may be a trypsinresistant protein. Significant reduction of CH activity occurred after incubation of bacteria with homologous antiserum, whereas little or no inhibition was found with heterologous or control serum. Prior heattreatment of cells at 56°C for 1 h resulted in significant reduction in CH activity, and the activity was completely lost after heating at 100°C for 1 h.
CH activity of bacterial components The OMP and LPS obtained from S.flexneri strain 2-61 1, S . dysenteriae type 1 strain 2-24623 and EIEC strain 2, either separately or combined, did not haemolyse RBCs in concentrations up to 10mg/ml (table 11). However, cells depleted of OMP showed
CONTACT HAEMOLYSIN OF EIEC AND SHIGELLAE
335
Table 11. Influence of RBC treatment and of bacterial treatment on CH activity of S . dysenteriae type 1, strain 2-24623, S._flexneristrain 2-61 1 and EIEC strain 2 CH activity ("A)*
Assay material Untreated bacteria and RBCs treated with lysozyme 250 pg/ml neuraminidase 100-1 000 U/ml N-acetyl neuraminic acid 10 mg/ml wheat germ agglutinin 1-10 pg/ml phytohaemagglutinin 0.1-1.0 pg/ml ConcanavalinA 0.01-0.1 pg/ml Untreated RBCs and bacteria treated with proteinase K 100 pg/ml trypsin 100 pg/ml homologous rabbit antiserum against S. dysenteriue S .Jlexneri EIEC antiserum against heterologous bacterial species normal rabbit antiserum heat: 56°C for 1 h 100°C for 1 h Untreated RBCs and bacteria depleted of OMP LPS Untreated RBCs and bacterial components : OMP 1-10 mg/ml LPS 1-10 mg/ml LPS + OMP 10-20 mg/ml
25-44 85-90
80-85 33-50 100 100 0
100
60
10
55
91-100 100
20 0
16-30 100 0 0 0
* Percentage of CH activity of positive control (untreated bacteria or untreated RBCs). 70-84% reduction in CH activity whereas cells depleted of LPS showed no loss. These results suggest that the CH molecule may be a component of bacterial outer membrane.
Discussion The best medium for demonstrating CH activity of EIEC was CYE broth (fig. 1); the higher amount of amino acids in this medium compared with the other two media, TSB and MM, may have influenced the CH activity.20 On the other hand, Shigella spp. exhibited maximal CH activity when grown in TSB. Other studies20921 showed that the capacity of S . dysenteriae to agglutinate RBCs, and the adherence of S . Jexneri type 2a to cultured mammalian cells, were maximal when grown in CYE; but this medium was not appropriate for CH activity of Shigella spp. However, other investigators6. found TSB suitable for CH activity and for invasion of epithelial cells in vitro by shigellae. These varied results probably indicate that the nutritional requirements for the expression of virulence factors by Shigella spp. and EIEC are diverse. All the bacterial species in the present study showed the highest CH activity when tested with guinea-pig RBCs (fig. 2), suggesting varied expression of CH receptor among different animal species. The lower susceptibilityof sheep RBCs may explain the absence of CH activity in S. boydii and EIEC as observed by Clerc et In general, CH production was favoured at slightly
alkaline pH (fig. 3), which was also optimal for growth. The relatively high CH activity of S . boydii at pH 6.0, and the sharp decrease in CH activity of EIEC at acidic or alkaline pH, indicate that the optimal pH for bacterial growth and for expression of CH activity are not the same with all bacteria. CH activity was completely destroyed by proteinase K (table 11), indicating its proteinaceous nature; but it was insensitive to trypsin. This may be because proteinase K is a mixture of endopeptidase and exopeptidase including carboxypeptidases and aminopeptidases with non-specific protease activity, whereas trypsin is highly selective in action, cleaving only those peptide bonds whose carboxyl function is donated by either a lysine or an arginine molecule.22 Thus, resistance to trypsin does not disprove that CH might be a protein molecule. Reduction in CH activity by pre-treatment of the bacteria with homologous antisera, and minimal inhibition with heterologous antisera, indicate the probable cell-surface location of the CH molecule which is also antigenic. Bacteria depleted of OMP, but not those depleted of LPS, showed significant decrease in CH activity, indicating its outer membrane location. On the other hand, isolated OMP and LPS, either separately or together, failed to show CH activity. If it is assumed that CH is a protein in the outer membrane, then inability of the isolated OMP to cause haemolysis needs to be explained. Live metabolically active bacteria are necessary for invasion;2 we also observed that cells from the logarithmic phase of growth were more efficient in CH assay than cells from the
336
K. HAIDER ET AL.
stationary phase, and little CH activity was detected when bacteria were either incubated with RBCs at 4°C or pre-heated at 56°C for 1 h. Thus, isolated cell components which include a relevant molecule may be non-functional in the absence of metabolically active bacteria. Decrease in CH activity after treatment of RBCs with lysozyme suggests that the CH molecule interacts with a lysozyme-sensitivemoiety on the RBC surface. Treatment with wheat germ agglutinin, specific for the lysozyme substrate of N-acetyl-D-glucosamine oligomers of three units in length (N,N,N"-triacetylc h i t ~ t r i o s e )also , ~ ~caused significant reduction in CH activity. On the other hand, treatment with phytohaemagglutinin and Concanavalin A, which do not interact with the lysozyme substrate chitotriose, failed to reduce CH activity. The minimal inhibitory effect of neuraminidase and NANA on CH activity, probably by non-specific interaction, renders the involvement of sialic acid unlikely. All this evidence suggests a chitotriose-like moiety as the probable receptor for CH. Studies with bacteria harbouring an intact plasmid and a modified plasmid, and with plasmidless derivatives, showed a correlation between CH activity and HeLa cell invasivenesse6Similar results were obtained in the present study, with two exceptions (table I) : avirulent S. dysenteriae type 1 strains had relatively high CH activity but were non-invasive, and virulent EIEC (0124) had low CH activity but was invasive. The reason for this lack of correlation is not apparent; but it is noticeable that, within each species, virulent strains always had higher CH activity and invasion potential than their avirulent counterparts. Our results showed that certain EIEC strains, especially 028, were even more invasive than those of S.$exneri and S. dysenteriae, and S. boydii was moderately invasive.
Other investigators also reported the invasive nature of EIEC in HEp-2 ~ e l l s . EIEC ~~,~ (028) ~ has also been found to possess high cell-surface hydrophobicity, similar to S . dysenteriae type 1 (unpublished observation). The absence of CH activity in EIEC and S . boydii noted by Clerc et al.7 could be due to their use of sheep RBCs (see above) or of different strains of bacteria. Reduced CH activity by bacterial cells grown at 30"C, and partial restoration after changing from 30°C to 37"C, indicate that the activity might be thermoregulated, as are other virulence properties of Shigella spp.-invasion of epithelial cells in vitro,6* expression of peptides involved in invasion,' binding of the dye Congo red, l 9 and haemagglutination.20 The CH activity in S.Jlexneri has been shown to be critical for lysis of the phagocytic vacuole after invasion of HeLa cells, which permits efficient intracellular multiplication.6 In this study, we have demonstrated CH activity and invasion of HeLa cells by EIEC; it is possible that CH activity of EIEC may also be essential for intracellular multiplication.
References
8. Sansonetti PJ, D'Hauleville H, Ecobichon C, Pourcel C. Molecular comparison of virulence plasmids in Shigella and enteroinvasive Escherichia coli. Ann Microbiol (Inst Pasteur) 1983; 134A:295-318. 9. Phl T, Phcsa S, Emijdy L, Voros S. Antigenic relationship among virulent enteroinvasive Escherichia coli, Shigella jlexneri, and Shigella sonnei detected by ELISA. Lancet 1983; 2: 102. 10. Ahmed ZU, Sarker MR, Sack DA. Nutritional requirements of shigellae for growth in a minimal medium. Infect Immun 1988; 56: 1007-1009. 1. Birnboim HC, Doly J. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res 1979; 7:1513-1523. 2. Meyers JA, Sanchez D, Elwell LP, Falkow S. Simple agarose gel electrophoretic method for the identification and characterization of plasmid deoxyribonucleic acid. J Bucterioll976; 127:1529-1537. 3. Sereny B. Experimental Shigella keratoconjunctivitis. Actu Microbiol Hung 1955;2 : 293-296. 14. Haider K, Azad AK, Qadri F, Nahar S, Ciznar 1. Role of plasmids in virulence-associated attributes and in 0antigen expression in Shigella dysenteriae type 1 strains. J Med Microbioll990; 33: 1-9. 15. Westphal 0, Jann K. Bacterial lipopolysaccharides : extraction with phenol-water and further applications of the procedure. In : Whistler RL (ed) General polysaccharides (Methods of carbohydrate chemistry, No. 5). New York, Academic Press. 1965: 83-91.
1. Formal SB, Hale TL, Sansonetti PJ. Invasive enteric pathogens. Rev Infect Dis 1983; 5: S702-S707. 2. Hale TL, Bonventre PF. Shigella infection of Henle intestinal epithelial cells: role of the bacterium. Infect Immun 1979; 24: 879-886. 3. Maurelli AT, Sansonetti PJ. Identification of a chromosomal gene controlling temperature-regulated expression of Shigella virulence. Proc Natl Acad Sci USA 1988; 85: 28202824. 4. Levine MM, Nataro JP, Karch H et al. The diarrheal response of humans to some classic serotypes of enteropathogenic Escherichiu coli is dependent on a plasmid encoding an enteroinvasiveness factor. J Infect Dis 1985; 152: 550-559. 5. Hale TL, Oaks EV, Formal SB. Identification and antigenic characterization of virulence-associated, plasmid-coded proteins of Shigella spp. and enteroinvasive Escherichia coli. Infect Immun 1985;50 : 620-629. 6. Sansonetti PJ, Ryter A, Clerc P, Maurelli AT, Mounier J. Multiplication of Shigellaflexneri within HeLa cells : lysis of the phagocytic vacuole and plasmid-mediated contact hemolysis. Infect Immun 1986; 51 : 461-469. 7. Clerc P, Baudry B, Sansonetti PJ. Plasmid-mediated contact haemolytic activity in Shigella species: correlation with penetration into HeLa cells. Ann Inst Pasteur Microbiol 1986; 137A:267-278.
This research was supported by the International Centre for Diarrhoea1 Disease Research, Bangladesh, which is supported by countries and agencies that share its concern about the impact of diarrhoea1 diseases in the developing countries. Current major donors are :The Aga Khan Foundation, Arab Gulf Fund, Australia, Bangladesh, Belgium, Canadian International Development Agency, Canadian International Development Research Centre, Danish International Development Agency, France, the Ford Foundation, Japan, the Netherlands, Norwegian Agency for International Development, Sweden, Swiss Development Cooperation, United Kingdom, United Nations Development Programme, United Nations Children's Fund, United Nations Capital Development Fund, United States Agency for International Development, World Health Organization and World University Service of Canada. We thank Dr F. Qadri for LPS of S. flexneri strain 2-61 1 and rabbit antiserum raised against this strain, and Mr M. Haque for secretarial assistance.
CONTACT HAEMOLYSIN OF EIEC AND SHIGELLAE 16. Oaks EV, Hale TL, Formal SB. Serum immune response to Shigella protein antigens in Rhesus monkeys and humans infected with Shigella spp. Infect Immun 1986;53: 57-63. 17. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976; 72 : 248-254. 18. Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. Colorimetric method for determination of sugars and related substances. Anal Chem 1956; 28: 350-356. 19. Qadri F, Hossain SA, Ciinir I et al. Congo red binding and salt aggregation as indicators of virulence in Shigella species. J Clin Microbioll988; 26: 1343-1348. 20. Qadri F, Haq S, Ciinir I. Hemagglutinating properties of Shigella dysenteriae type 1 and other Shigella species. Infect Immun 1989;57 : 2909-29 1 1 . 21. Izhar M, Nuchamowitz Y, Mirelman D. Adherence of Shigella
22. 23.
24.
25.
337
flexneri to guinea pig intestinal cells is mediated by a mucosal adhesin. Infect Immun 1982; 35: 1 1 10-1 118. Lehninger AL. Biochemistry; the molecular basis of cell structure and function. New York, Worth Publishers, Inc. 1975 : 1104. Keusch GT, Jacewicz M. Pathogenesis of Shigella diarrhea. VII. Evidence for a cell membrane toxin receptor involving Pl44-linked N-acetyl-D-glucosamine oligomers. J Exp Med 1977; 146; 535-546. Small PLC, Isberg RR, Falkow S. Comparison of the ability of enteroinvasive Escherichia coli, Salmonella typhimurium, Yersinia pseudotuberculosis and Yersinia enterocolitica to enter and replicate within HEp-2 cells. Infect Immun 1987; 55: 1674-1679. Falkow S, Small P, Isberg R, Hayes SF, Corwin D. A molecular strategy for the study of bacterial invasion. Rev Infect Dis 1987; 9 Suppl: S450-S455.
01991 The Pathological Society of Great Britain and Ireland
Contact- haemolysin production by entero-invasive Escherichia cok and shigellae KHALEDA HAIDER, M. J. ALBERT, A. HOSSAIN* and S. NAHAR Laboratory Sciences Division, International Centre for Diarrhoea1 Disease Research, GPO Box 128, Dhaka 1000, and Department of Microbiology, University of Dhaka, Dhaka 1000, Bangladesh
Summary. Entero-invasive Escherichia coli (EIEC) and shigellae were tested for contacthaemolysin (CH) with red blood cells (RBCs) of guinea-pig, rabbit, rat, mouse, monkey, man, sheep and chicken; all bacteria showed the best lysis with guinea-pig RBCs. The best culture medium for CH activity of shigellae was tryptic soy broth, and for EIEC it was casamino acid-yeast extract broth with 1 mM CaC12.CH production by all species was best at the slightly alkaline pH which is optimal for growth; it was also dependent on the presence of a large (140-Mda) plasmid. Pre-treatment of bacteria with homologous antisera inhibited CH activity. Various treatments of bacterial cells and RBCs suggested that CH may be a protein molecule, and that a chitotriose-like moiety may serve as CH receptor. RBCs that were incubated with bacteria at 4"C, or with heat-killed bacteria at 37"C, were not lysed; also, isolated cell-surface components (lipopolysaccharideand outer-membrane protein) did not lyse RBCs. This suggests that metabolically active cells are required for CH activity. Production of CH by both EIEC and shigellae is consistent with a common mechanism for the virulence of these organisms.
Introduction
The large virulence-plasmidsof EIEC and all species of Shigella have extensive homologous sequences and are functionally exchangeable.' As the pathogenesis of EIEC and Shigella spp. is similar, this study was initiated to investigate whether CH was also produced by S . boydii and EIEC, other species of Shigella being included for comparison. Factors influencing the production and activity of CH were investigated by the use of virulent and avirulent pairs. Preliminary experiments were also performed to study the chemical nature of CH.
Disease caused by shigellae involves a complex process of invasion and multiplication in the mucosa of the human colon. The process of invasion requires entry of the bacterium into the enterocytes, and lysis of the membranous vacuole which surrounds it after invasion, to allow free multiplication of the pathogen within the cytoplasm. This leads to spread of the organisms to the lamina propria, triggering a severe inflammatory response in the gut mucosa with abscesses and ~ l c e r a t i o n . ~All ~ these features are regulated by a large non-self-conjugative plasmid of 120-140 Mda called the virulence plasmid. Enteroinvasive Escherichia coli (EIEC) produces dysentery like that caused by shigellae; it also possesses a 140Mda plasmid which is associated with vir~lence.~. Sansonetti et a1.6 described a virulence-plasmidmediated, thermoregulated contact-haemolysin (CH) in Shigella, which was active against sheep red blood cells (SRBC). Close contact was required between virulent bacteria and SRBC which was achieved by centrifugation. The production of CH correlated with its capacity to destroy the membranous vacuole surrounding the bacteria after invasion of HeLa cells, and with efficient intracellular multiplication. CH was also demonstrated in S. dysenteriae and S . sonnei strains in another s t ~ d y . ~
Materials and methods Bacterial strains and culture conditions The bacterial strains are shown in table I. Dr D. N. Taylor, Armed Forces Research Institute of Medical Sciences (AFRIMS), Bangkok, kindly provided 26 strains of EIEC from patients with invasive diarrhoea in Thailand. Five similar isolates were from central Australia (2), Bangladesh (2) and Hungary (l).9 The avirulent EIEC strain 2/33 is an isogenic mutant strain of virulent EIEC strain 2, isolated in Hungary ;9 these were kindly supplied by Dr P. Echeverria (AFRIMS, Bangkok). E. coli strain 36 000 was obtained from the Centers for Disease Control, Atlanta, GA, USA. All Shigella strains were isolated from patients at the Clinical Research Centre, International Centre for Diarrhoeal Disease Research, Bangladesh. The CH
Received 19 Dec. 1990; revised version accepted 3 April 1991. 330
CONTACT HAEMOLYSIN OF EIEC AND SHIGELLAE
331
Table I. CH activity and HeLa-cell invasiveness of shigellae and EIEC Number of strains
CH activity (OD,,,)
Number* of bacteria/HeLa cell
Virulent Avirulent Rough
5 5 4
2.19 0.89 0.12
53 0 0
S . sonnei
Virulent Avirulent Rough
5 5 4
1.65 0.12 0.15
54 0 0
S . boydii
Virulent Avirulent
5 5
1.16 0.34
11 0
S.Jlexneri l a lb 2a 2b 3a 3b 4a 5 Y 6 2a
Virulent Virulent Virulent Virulent Virulent Virulent Virulent Virulent Virulent Virulent Avirulent
2 2 5 2 2 2 1 2 2 2 5
1a06 0.72 2.43 0.7 1 1.10 1.14 1.10 2.00 2.27 1.44 0.5 1
25 5 70 3 15 15 12 51 65 20 0
EIEC
Virulent Virulent Virulent Virulent Virulent Virulent Virulent Avirulent
24 2 1 1 1 1 1 1
2.72 1.35 0.87 0.82 0.90 0.56 1.01 0.50
Avirulent
1
0.10
152 15 3 6 6 4 27 0 0
Organism S . dysenteriae type 1
028 029 0143 0144 0164 0124 0136 0143
E.coli 36000
* Each of these numbers is the approximate mean of results from at least three experiments. Virulent strains gave a positive Sereny result, invaded HeLa-cells and bound Congo red dye; avirulent strains did none of these. Virulent strains had a 120-140-Mda virulence plasmid; avirulent strains had either no virulence plasmid or a deleted one, Rough strains autoagglutinated in physiological saline and did not have the virulence plasmid.
activity of bacterial components was studied with S . JEexneri strain 2-61 1, S. dysenteriae type 1 strain Z24623, and EIEC strain 2. All the strains were stored at - 70°C in Trypticase Soy Broth containing yeast extract 0.6% w/v (TSB; Gibco) with glycerine 15%v/v. The growth media were TSB, Casamino Acid-Yeast Extract Broth (Difco) supplemented with 1 mM calcium chloride (CYE) and minimal medium (MM) as described by A hmed et al.
Plasmid DNA was isolated by the method of Birnboim and Doly' and electrophoresed in agarose 0.7% gel.
Assay of CH
Testsfor bacterial virulence
CH activity was measured by the method of Sansonetti et a1.6 Briefly, bacteria were washed once and suspended in phosphate-buffered saline pH 7.3 (PBS) to a concentration of 2 x lo1' cfu/ml; blood was collected in Alsever's solution, washed twice, and the packed RBCs were resuspended in PBS to a concentration of 4 x lo9 RBCs/ml; 5O-pl volumes of bacterial suspension were mixed with equal volumes of RBCs in 96-well microtitration plates, and centrifuged at 2200 g for 10 min to achieve close contact between bacteria and RBCs. The plates were incubated at
Each strain was tested for virulence, both by invasion of HeLa cell monolayers6and by the Sereny test.I3 The latter was performed by inoculating 20 pl of bacterial suspension containing 3 x lo8 cfu/ml into the conjunctival sac of an adult guinea-pig; strains producing keratoconjunctivitis within 72 h were considered virulent. For the former, non-confluent monolayers of HeLa cells on 12-mm cover glass (Kimble, Toledo, OH, USA) in a vial were inoculated with bacteria, centrifuged for 10 min at 2200 g and incubated for 2 h
either 37°C or 4°C for 2 h. Pellets were resuspended in cold PBS, centrifuged at 4°C for 15 min and the OD,,o of the supernates was measured with an ELISA reader (Titertek-Multiscan, Flow Laboratories).
Plasmid analysis
332
K. HAIDER ET AL.
at 37°C. Non-invasive bacteria were washed off with PBS. Monolayers were fixed in methanol and stained with Giemsa dye, and the percentage of infected cells was determined by light-microscopy.To calculate the average number of viable bacteria in each infected HeLa cell, strains were inoculated on to monolayers of HeLa cells in a vial, centrifuged for 10 min at 2200 g, incubated for 15 min, washed three times with PBS, covered with 2ml of Minimum Essential Medium (Gibco) containing gentamicin 200 pg/ml, and reincubated for 2 h at 37"C, Monolayers were then washed, cells were trypsinised and counted, and finally cells were lysed with a solution of sodium deoxycholate 0.5% in distilled water. Dilutions of this suspension were plated to count the number of bacteria/vial. The average number of bacteria in each infected HeLa cell in a vial was calculated as follows : number of viable bacteria number of HeLa cells x percentage of infected HeLa cells
For comparison of the invasiveness of different organisms, the chosen multiplicity of infection was 100 bacteria/HeLa cell. Each strain was tested at least three times; results are presented as the mean from three experiments.
Treatment of erythrocytes Erythrocytes (RBCs) were treated at 37°C for 1 h with various reagents (Sigma): N-acetylneuraminic acid 1% w/v (NANA), lysozyme250 pg/ml, neuraminidase 100-1000 U/ml, and lectins-wheat germ agglutinin 10 pg/ml, phytohaemagglutinin 1.0 pg/ml and Concanavalin A 0.1 pg/ml. Higher concentrations of phytohaemagglutinin or Concanavalin A caused lysis of guinea-pig RBCs.
Treatment of bacteria Bacteria (10gcfu)were treated with: (1) enzymes for 4 h at 37°C-trypsin (Flow Laboratories, Herts) 100 pg/ml, proteinase K (Sigma) 100 pg/ml; (2) rabbit sera for 1 h at 37°C-normal serum and homologous and heterologous antisera against whole bacterial cells;14 (3) heat, at 56°C or 100°C for 1 h.
Extraction of bacterial components Lipopolysaccharide (LPS) of S . JEexneri strain Z61 1, S . dysenteriae type 1 strain 2-24623, and EIEC strain 2 were extracted with hot phenol-water" and purified by ultracentrifugation at 100 000 g for 4 h. Nucleic acid and protein were removed by treatment with DNAase and RNAase, and proteinase K 100 pg/ ml, respectively. Bacterial cells were depleted of LPS by treating cells with edetic acid. Outer-membrane protein (OMP) was prepared by water extraction. LPS and OMP were freeze-dried and reconstituted in PBS. Protein was estimated by d~e-binding'~ and
'
carbohydrate in LPS was estimated by the method of Dubois et aZ.'* CH activity of LPS and OMP was tested at concentrations up to 10 mg/ml. Bacterial cells depleted of OMP and LPS were also assayed for CH activity.
Results InJtuence of growth medium on CH activity Bacteria were grown with shaking in TSB, CYE and MM, and their CH activity was assayed with sheep RBCs (fig. 1). For S.JEexneri,TSB was the best medium, whereas S . dysenteriae and EIEC exhibited maximal activity when grown in CYE; on the other hand, CYE provided only poor CH activity by S . Jlexneri, S. sonnei and S. boydii. CH activity was reduced or undetectable when EIEC and shigellae were grown in static broth or on agar plates of these media.
CH activity on RBCs of different animal species RBCs from guinea-pig, rabbit, rat, mouse, monkey, man (blood group 0), sheep and chicken were compared (fig. 2). All the bacterial species exhibited maximal CH activity against guinea-pig RBCs ;hence, the latter were used for subsequent experiments. Except with S . JEexneri and EIEC, sheep RBCs were poorly lysed. The influence of growth medium was re-tested with guinea-pig RBCs : TSB was the most suitable medium for production of CH by shigellae and CYE for EIEC (fig. 1).
InJtuence ofpH of growth medium on CH activity For S.JEexneri, S . dysenteriae type 1 and S . sonnei, neutral or slightly alkaline pH (7.3-8.5) was suitable for CH production. However, with S . boydii, the activity was reduced as the pH was increased from 7-3 to 9.0, and there was not much reduction in CH activity as the pH decreased from 7-3 to 6.0. EIEC showed a marked optimum at pH 7.3, and this pH was also near the optimal for shigellae (fig. 3).
Inzuence of duration and temperature of growth on CH activity CH activity was examined in bacterial cultures grown for 4-48 h. It was maximal at 6-10 h, but beyond 24 h bacteria exhibited little haemolytic activity. CH activity of both shigellae and EIEC was greatly dependent on growth temperature : strains which were strongly positive when grown at 37°C showed reduced activity when grown at 30°C. Activity could be partially restored by changing the growth temperature to 37°C for 2 h before incubation with RBCs, but
CONTACT HAEMOLYSIN OF EIEC AND SHIGELLAE
333
Fig. 1. Influence of growth medium on CH activity of ShigeZZa spp. and EIEC. Bacteria grown in TSB (m), CYE (R) and MM (m) were tested with sheep RBC. Blank bars (0) to the right of the marked bars indicate CH activity of bacteria grown in the same medium but tested with guinea-pig RBC. OD,,o was measured with an ELISA reader. S. flex = S.jexneri; S. dys 1 = S . dysenteriae type 1.
Fig. 2. CH activity of ShigelZa spp. and EIEC against RBCs of different animal species: guinea-pig, monkey, ;man, W ;sheep, ; chicken, @ and see footnote to fig. 1.
extension of this period to 4 h did not result in further increase of CH activity. No CH activity was shown when the assay was performed at 4"C, even with bacteria grown optimally at 37°C.
.;
rabbit,
m; rat, 1; mouse, @ ;
Association of large virulenceplasmid with C H activity Plasmid profiles of several virulent strains, both shigellaeand EIEC, and of their avirulent counterparts
334
K. HAIDER ET AL.
3-0
2-5
2.0
1.5
1.0
0-5
0 S.flex
S.dys 1
S.sonnei
S.boydii
EIEC
Fig. 3. Influence of pH of growth medium on CH activity of Shigella spp. and EIEC : pH 6.0, TJ ;6.5, IS!;7.3, see footnote to fig. 1.
were analysed to study association of the 120-140Mda plasmid with CH activity. Avirulence was determined by the inability to bind the dye Congored, to produce keratoconjunctivitis in the guineapig,I3 and to invade HeLa cells.6 In all the species, CH activity appeared to be plasmid mediated, because loss of the large plasmid, or reduction in size by deletion, resulted in almost complete absence of CH activity (table I).
''
Correlationof CH activity with invasion potential of bacteria S . flexneri serotypes 2a, 5 and Y, S . dysenteriae type 1, and S . sonnei showed similar invasion potential after incubation for 2 h, reaching 51-70 bacteria/ HeLa-cell (table I), However, EIEC type 0 2 8 showed a much higher degree of invasion (152 bacteria/cell) than shigellae. Other serotypes of EIEC and S.flexneri and strains of S . boydii showed lower invasion capability (3-27 bacteria/cell). Generally, CH activity correlated well with the invasion potential of the organism. However, avirulent strains of S. dysenteriae type 1, although non-invasive in HeLa-cell assay, exhibited slightly higher CH activity than some virulent strains of other ShigeZZa spp. and EIEC. Also, one virulent EIEC strain ( 0 124) exhibited low CH activity although it was moderately invasive (table I).
;8.0,
; 8.5, El;9.0, Ei and
Influence of various treatments of RBCs and bacteria on CH activity Pre-treatment of RBCs with lysozyme resulted in significant reduction (56-75%) in CH activity, but neuraminidase and N ANA had minimal inhibitory effect (table 11). Of the three lectins, wheat germ agglutinin showed 50-67% inhibition of CH activity, whereas phytohaemagglutinin and Concanavalin A had no effect. No reduction in CH activity was noted when bacteria were treated with trypsin, but CH activity was completely lost after treatment with proteinase K, indicating that the CH molecule may be a trypsinresistant protein. Significant reduction of CH activity occurred after incubation of bacteria with homologous antiserum, whereas little or no inhibition was found with heterologous or control serum. Prior heattreatment of cells at 56°C for 1 h resulted in significant reduction in CH activity, and the activity was completely lost after heating at 100°C for 1 h.
CH activity of bacterial components The OMP and LPS obtained from S.flexneri strain 2-61 1, S . dysenteriae type 1 strain 2-24623 and EIEC strain 2, either separately or combined, did not haemolyse RBCs in concentrations up to 10mg/ml (table 11). However, cells depleted of OMP showed
CONTACT HAEMOLYSIN OF EIEC AND SHIGELLAE
335
Table 11. Influence of RBC treatment and of bacterial treatment on CH activity of S . dysenteriae type 1, strain 2-24623, S._flexneristrain 2-61 1 and EIEC strain 2 CH activity ("A)*
Assay material Untreated bacteria and RBCs treated with lysozyme 250 pg/ml neuraminidase 100-1 000 U/ml N-acetyl neuraminic acid 10 mg/ml wheat germ agglutinin 1-10 pg/ml phytohaemagglutinin 0.1-1.0 pg/ml ConcanavalinA 0.01-0.1 pg/ml Untreated RBCs and bacteria treated with proteinase K 100 pg/ml trypsin 100 pg/ml homologous rabbit antiserum against S. dysenteriue S .Jlexneri EIEC antiserum against heterologous bacterial species normal rabbit antiserum heat: 56°C for 1 h 100°C for 1 h Untreated RBCs and bacteria depleted of OMP LPS Untreated RBCs and bacterial components : OMP 1-10 mg/ml LPS 1-10 mg/ml LPS + OMP 10-20 mg/ml
25-44 85-90
80-85 33-50 100 100 0
100
60
10
55
91-100 100
20 0
16-30 100 0 0 0
* Percentage of CH activity of positive control (untreated bacteria or untreated RBCs). 70-84% reduction in CH activity whereas cells depleted of LPS showed no loss. These results suggest that the CH molecule may be a component of bacterial outer membrane.
Discussion The best medium for demonstrating CH activity of EIEC was CYE broth (fig. 1); the higher amount of amino acids in this medium compared with the other two media, TSB and MM, may have influenced the CH activity.20 On the other hand, Shigella spp. exhibited maximal CH activity when grown in TSB. Other studies20921 showed that the capacity of S . dysenteriae to agglutinate RBCs, and the adherence of S . Jexneri type 2a to cultured mammalian cells, were maximal when grown in CYE; but this medium was not appropriate for CH activity of Shigella spp. However, other investigators6. found TSB suitable for CH activity and for invasion of epithelial cells in vitro by shigellae. These varied results probably indicate that the nutritional requirements for the expression of virulence factors by Shigella spp. and EIEC are diverse. All the bacterial species in the present study showed the highest CH activity when tested with guinea-pig RBCs (fig. 2), suggesting varied expression of CH receptor among different animal species. The lower susceptibilityof sheep RBCs may explain the absence of CH activity in S. boydii and EIEC as observed by Clerc et In general, CH production was favoured at slightly
alkaline pH (fig. 3), which was also optimal for growth. The relatively high CH activity of S . boydii at pH 6.0, and the sharp decrease in CH activity of EIEC at acidic or alkaline pH, indicate that the optimal pH for bacterial growth and for expression of CH activity are not the same with all bacteria. CH activity was completely destroyed by proteinase K (table 11), indicating its proteinaceous nature; but it was insensitive to trypsin. This may be because proteinase K is a mixture of endopeptidase and exopeptidase including carboxypeptidases and aminopeptidases with non-specific protease activity, whereas trypsin is highly selective in action, cleaving only those peptide bonds whose carboxyl function is donated by either a lysine or an arginine molecule.22 Thus, resistance to trypsin does not disprove that CH might be a protein molecule. Reduction in CH activity by pre-treatment of the bacteria with homologous antisera, and minimal inhibition with heterologous antisera, indicate the probable cell-surface location of the CH molecule which is also antigenic. Bacteria depleted of OMP, but not those depleted of LPS, showed significant decrease in CH activity, indicating its outer membrane location. On the other hand, isolated OMP and LPS, either separately or together, failed to show CH activity. If it is assumed that CH is a protein in the outer membrane, then inability of the isolated OMP to cause haemolysis needs to be explained. Live metabolically active bacteria are necessary for invasion;2 we also observed that cells from the logarithmic phase of growth were more efficient in CH assay than cells from the
336
K. HAIDER ET AL.
stationary phase, and little CH activity was detected when bacteria were either incubated with RBCs at 4°C or pre-heated at 56°C for 1 h. Thus, isolated cell components which include a relevant molecule may be non-functional in the absence of metabolically active bacteria. Decrease in CH activity after treatment of RBCs with lysozyme suggests that the CH molecule interacts with a lysozyme-sensitivemoiety on the RBC surface. Treatment with wheat germ agglutinin, specific for the lysozyme substrate of N-acetyl-D-glucosamine oligomers of three units in length (N,N,N"-triacetylc h i t ~ t r i o s e )also , ~ ~caused significant reduction in CH activity. On the other hand, treatment with phytohaemagglutinin and Concanavalin A, which do not interact with the lysozyme substrate chitotriose, failed to reduce CH activity. The minimal inhibitory effect of neuraminidase and NANA on CH activity, probably by non-specific interaction, renders the involvement of sialic acid unlikely. All this evidence suggests a chitotriose-like moiety as the probable receptor for CH. Studies with bacteria harbouring an intact plasmid and a modified plasmid, and with plasmidless derivatives, showed a correlation between CH activity and HeLa cell invasivenesse6Similar results were obtained in the present study, with two exceptions (table I) : avirulent S. dysenteriae type 1 strains had relatively high CH activity but were non-invasive, and virulent EIEC (0124) had low CH activity but was invasive. The reason for this lack of correlation is not apparent; but it is noticeable that, within each species, virulent strains always had higher CH activity and invasion potential than their avirulent counterparts. Our results showed that certain EIEC strains, especially 028, were even more invasive than those of S.$exneri and S. dysenteriae, and S. boydii was moderately invasive.
Other investigators also reported the invasive nature of EIEC in HEp-2 ~ e l l s . EIEC ~~,~ (028) ~ has also been found to possess high cell-surface hydrophobicity, similar to S . dysenteriae type 1 (unpublished observation). The absence of CH activity in EIEC and S . boydii noted by Clerc et al.7 could be due to their use of sheep RBCs (see above) or of different strains of bacteria. Reduced CH activity by bacterial cells grown at 30"C, and partial restoration after changing from 30°C to 37"C, indicate that the activity might be thermoregulated, as are other virulence properties of Shigella spp.-invasion of epithelial cells in vitro,6* expression of peptides involved in invasion,' binding of the dye Congo red, l 9 and haemagglutination.20 The CH activity in S.Jlexneri has been shown to be critical for lysis of the phagocytic vacuole after invasion of HeLa cells, which permits efficient intracellular multiplication.6 In this study, we have demonstrated CH activity and invasion of HeLa cells by EIEC; it is possible that CH activity of EIEC may also be essential for intracellular multiplication.
References
8. Sansonetti PJ, D'Hauleville H, Ecobichon C, Pourcel C. Molecular comparison of virulence plasmids in Shigella and enteroinvasive Escherichia coli. Ann Microbiol (Inst Pasteur) 1983; 134A:295-318. 9. Phl T, Phcsa S, Emijdy L, Voros S. Antigenic relationship among virulent enteroinvasive Escherichia coli, Shigella jlexneri, and Shigella sonnei detected by ELISA. Lancet 1983; 2: 102. 10. Ahmed ZU, Sarker MR, Sack DA. Nutritional requirements of shigellae for growth in a minimal medium. Infect Immun 1988; 56: 1007-1009. 1. Birnboim HC, Doly J. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res 1979; 7:1513-1523. 2. Meyers JA, Sanchez D, Elwell LP, Falkow S. Simple agarose gel electrophoretic method for the identification and characterization of plasmid deoxyribonucleic acid. J Bucterioll976; 127:1529-1537. 3. Sereny B. Experimental Shigella keratoconjunctivitis. Actu Microbiol Hung 1955;2 : 293-296. 14. Haider K, Azad AK, Qadri F, Nahar S, Ciznar 1. Role of plasmids in virulence-associated attributes and in 0antigen expression in Shigella dysenteriae type 1 strains. J Med Microbioll990; 33: 1-9. 15. Westphal 0, Jann K. Bacterial lipopolysaccharides : extraction with phenol-water and further applications of the procedure. In : Whistler RL (ed) General polysaccharides (Methods of carbohydrate chemistry, No. 5). New York, Academic Press. 1965: 83-91.
1. Formal SB, Hale TL, Sansonetti PJ. Invasive enteric pathogens. Rev Infect Dis 1983; 5: S702-S707. 2. Hale TL, Bonventre PF. Shigella infection of Henle intestinal epithelial cells: role of the bacterium. Infect Immun 1979; 24: 879-886. 3. Maurelli AT, Sansonetti PJ. Identification of a chromosomal gene controlling temperature-regulated expression of Shigella virulence. Proc Natl Acad Sci USA 1988; 85: 28202824. 4. Levine MM, Nataro JP, Karch H et al. The diarrheal response of humans to some classic serotypes of enteropathogenic Escherichiu coli is dependent on a plasmid encoding an enteroinvasiveness factor. J Infect Dis 1985; 152: 550-559. 5. Hale TL, Oaks EV, Formal SB. Identification and antigenic characterization of virulence-associated, plasmid-coded proteins of Shigella spp. and enteroinvasive Escherichia coli. Infect Immun 1985;50 : 620-629. 6. Sansonetti PJ, Ryter A, Clerc P, Maurelli AT, Mounier J. Multiplication of Shigellaflexneri within HeLa cells : lysis of the phagocytic vacuole and plasmid-mediated contact hemolysis. Infect Immun 1986; 51 : 461-469. 7. Clerc P, Baudry B, Sansonetti PJ. Plasmid-mediated contact haemolytic activity in Shigella species: correlation with penetration into HeLa cells. Ann Inst Pasteur Microbiol 1986; 137A:267-278.
This research was supported by the International Centre for Diarrhoea1 Disease Research, Bangladesh, which is supported by countries and agencies that share its concern about the impact of diarrhoea1 diseases in the developing countries. Current major donors are :The Aga Khan Foundation, Arab Gulf Fund, Australia, Bangladesh, Belgium, Canadian International Development Agency, Canadian International Development Research Centre, Danish International Development Agency, France, the Ford Foundation, Japan, the Netherlands, Norwegian Agency for International Development, Sweden, Swiss Development Cooperation, United Kingdom, United Nations Development Programme, United Nations Children's Fund, United Nations Capital Development Fund, United States Agency for International Development, World Health Organization and World University Service of Canada. We thank Dr F. Qadri for LPS of S. flexneri strain 2-61 1 and rabbit antiserum raised against this strain, and Mr M. Haque for secretarial assistance.
CONTACT HAEMOLYSIN OF EIEC AND SHIGELLAE 16. Oaks EV, Hale TL, Formal SB. Serum immune response to Shigella protein antigens in Rhesus monkeys and humans infected with Shigella spp. Infect Immun 1986;53: 57-63. 17. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976; 72 : 248-254. 18. Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. Colorimetric method for determination of sugars and related substances. Anal Chem 1956; 28: 350-356. 19. Qadri F, Hossain SA, Ciinir I et al. Congo red binding and salt aggregation as indicators of virulence in Shigella species. J Clin Microbioll988; 26: 1343-1348. 20. Qadri F, Haq S, Ciinir I. Hemagglutinating properties of Shigella dysenteriae type 1 and other Shigella species. Infect Immun 1989;57 : 2909-29 1 1 . 21. Izhar M, Nuchamowitz Y, Mirelman D. Adherence of Shigella
22. 23.
24.
25.
337
flexneri to guinea pig intestinal cells is mediated by a mucosal adhesin. Infect Immun 1982; 35: 1 1 10-1 118. Lehninger AL. Biochemistry; the molecular basis of cell structure and function. New York, Worth Publishers, Inc. 1975 : 1104. Keusch GT, Jacewicz M. Pathogenesis of Shigella diarrhea. VII. Evidence for a cell membrane toxin receptor involving Pl44-linked N-acetyl-D-glucosamine oligomers. J Exp Med 1977; 146; 535-546. Small PLC, Isberg RR, Falkow S. Comparison of the ability of enteroinvasive Escherichia coli, Salmonella typhimurium, Yersinia pseudotuberculosis and Yersinia enterocolitica to enter and replicate within HEp-2 cells. Infect Immun 1987; 55: 1674-1679. Falkow S, Small P, Isberg R, Hayes SF, Corwin D. A molecular strategy for the study of bacterial invasion. Rev Infect Dis 1987; 9 Suppl: S450-S455.
